Jorge:
My 2 cents:
1) Although there wasn't a lot of motion in terms of distance, the
realignment parameters do have a significant rhythmicity to them. It
looks like the person was shaking their head subtly giving rhythmic
motion involving mostly x translation and roll. This is very suspect.
You might look at a fourier transform of the motion parameters to
make sure that the frequency of maximum power is not overlapping your
task.
2) I'm not sure how you set up the analysis but the design matrix
shows only 2 conditions. SPM allowed you to perform a contrast on a
single condition so basically i think you're looking at which areas
of the brain show a signal different than 0. Which, as you can see is
most of the brain. In looking at your design matrix there seem to be
3 active periods and I don't think you modeled the rest at all. So
assuming your paradigm was
face fix house fix face fix house fix face fix house fix
here is a suggested setup.
sessions=1
scans=120
conditions =4 (you need a condition for face, for the face fixation,
for house and for the house fixation)
stochastic= no
Fixed|variable = fixed
onset of condition 1 (face) = 0
SOA of condition 1 =40 (scans to return to face condition)
onset of condition 2 (face fix) = 12.
SOA of condition 2 =40
onset of condition 3 (house) = 20
SOA of condition 3 = 40
onset of condition 4 (house fix) = 32
SOA of condition 4 = 40
length of condition 1 = 12
length of condition 2 = 8
length of condition 3 = 12
length of condition 4 = 8
parametric = no
volterra = no
confounds = no.
This should produce a design matrix with 4 conditions which you can
properly contrast.
So
face - fix = 1 -1 0 0
house - fix = 0 0 1 -1
face -house (each vs. its respective rest) 1 -1 -1 1
and so on.
hope this helps,
Darren
At 12:40 PM -0700 1/25/00, Jorge Jovicich wrote:
>Dear SPM Gurus,
>
>I am obtaining very strange results from the statistics analysis (SPM99b)
>of a simple fMRI experiment and would like to see if you can help me
finding
>what's wrong.
>
>To summarize: SPM finds VERY strong activation (t~10) in the *WHOLE* brain
>which is NOT due to motion whereas the raw image data gives no indication
>of this at all.
>
>The experiment was a face localiser task (passive viewing of faces and
>houses).
>The paridgm was: fixation-[Face-fixation-House-fixation]x3
>The epoch lenghts were fixation=20s and Face=House=30s, TR was 2.5s and
>pixel size was 3.125mm.
>Before the statistics analysis I performed the realigment (with the
>default options), Talairach normalization (also with defaults but to
>a pixel size
>[3x3x3]), and spatial smoothing (10x10x10).
>
>SPM shows me that:
>- there was no significant motion during the scanning session (translation
>< 0.4 mm, rotation < 0.3 degrees, see attachment motion.gif);
>- during the face presentation periods the occipital cortex was activated,
>as I would have expected (see attachment face.gif);
>- during the house presentation periods THE-WHOLE-BRAIN was significantly
>activated (see attachments house*.gif). This I find hard to believe if it
>is true that there was no motion.
>
>So far I have:
>-checked that the data in the configuration file is correct;
>-checked that the fitted time course during the house presentation period
>is really suggesting activation within the whole brain;
>-checked that I obtain the same results with SPM when I skip the motion
>correction;
>-looked at the un-realigned images in a movie-fashion and confirmed that
>no strong motion is visible;
>-looked at the time courses of the raw image data (i.e., no motion
>correction, no Talairach normalization, no smoothing) and confirmed
>that there is some
>activation in the occipital cortex during the stimuli presentation. But by
no
>means there is activation with t values close to 10 in the overall brain.
Such
>activation should be pretty clear, even in unsmoothed images...
>
>Do you have any ideas of what could be going on?
>I can send the raw image data if you'll be willing to have closer look at
>it.
>
>Thanks,
>Jorge
>
>-------------------------------------------------------------------
>Jorge Jovicich, Dr. rer. nat.
>
>Dept. of Neurology | Computation and Neural Systems
>Harbor-UCLA Medical Center | California Institute of Technology
>1124 W Carson St. B-4 | 1200 E California Blv., BI 139-74
>Torrance, CA 90502, USA | Pasadena, CA 91125, USA
>310-222-5656/7 | 626-395-2882
>310-222-5658 (fax) | 626-796-8876 (fax)
>[log in to unmask] | [log in to unmask]
>
>
>
>Content-id: <[log in to unmask]>
>Content-type: IMAGE/GIF; name="<JOVICICH>MOTION.GIF;1"
>Content-description:
>Content-disposition: attachment; filename="<JOVICICH>MOTION.GIF;1"
>
>Attachment converted: DRG:<JOVICICH>MOTION.GIF;1 (GIFf/JVWR) (00085F03)
>Content-id: <[log in to unmask]>
>Content-type: IMAGE/GIF; name="<JOVICICH>FACE.GIF;1"
>Content-description:
>Content-disposition: attachment; filename="<JOVICICH>FACE.GIF;1"
>
>Attachment converted: DRG:<JOVICICH>FACE.GIF;1 (GIFf/JVWR) (00085F04)
>Content-id: <[log in to unmask]>
>Content-type: IMAGE/GIF; name="<JOVICICH>HOUSE1.GIF;1"
>Content-description:
>Content-disposition: attachment; filename="<JOVICICH>HOUSE1.GIF;1"
>
>Attachment converted: DRG:<JOVICICH>HOUSE1.GIF;1 (GIFf/JVWR) (00085F05)
>Content-id: <[log in to unmask]>
>Content-type: IMAGE/GIF; name="<JOVICICH>HOUSE2.GIF;1"
>Content-description:
>Content-disposition: attachment; filename="<JOVICICH>HOUSE2.GIF;1"
>
>Attachment converted: DRG:<JOVICICH>HOUSE2.GIF;1 (GIFf/JVWR) (00085F06)
>Content-id: <[log in to unmask]>
>Content-type: IMAGE/GIF; name="<JOVICICH>HOUSE3.GIF;1"
>Content-description:
>Content-disposition: attachment; filename="<JOVICICH>HOUSE3.GIF;1"
>
>Attachment converted: DRG:<JOVICICH>HOUSE3.GIF;1 (GIFf/JVWR) (00085F07)
Darren R. Gitelman, M.D.
-------------------------------------------------------------------------
Cognitive Neurology and Alzheimer's Disease Center
E-mail: [log in to unmask] WWW: http://www.brain.nwu.edu
Voice: (312) 908-9023 Fax: (312) 908-8789
Northwestern Univ., 320 E. Superior St., Searle 11-470, Chicago, IL 60611
-------------------------------------------------------------------------
|